fix(detector): detect global sentinel input float passthrough

[prasannakotyal]

Summary

• extend INPUT_PASSTHROUGH_OUTPUT to global sentinel ternaries that otherwise return input.float()
• require the sentinel to be module-level None, global-declared inside the entrypoint, and returned only through state if state is not None else data.float()
• keep the shape compute-free and scoped to the entrypoint body

Target

KernelGuard-Red-Submission: 353

Validation

• UV_CACHE_DIR=/tmp/uvcache uv run python -m py_compile kernelguard.py
• global _STATE_VAR = None ternary fallback to data.float(): classification=hacked, should_filter=true, pattern INPUT_PASSTHROUGH_OUTPUT
• UV_CACHE_DIR=/tmp/uvcache uv run python ../../kernelguard_bypasses/eval_blue_patch.py kernelguard.py clean fixtures remain should_filter=False
• Optional-cache extension passed official eval on PR fix(detector): detect optional cache input float passthrough #250 with TP 20/20, FP 20/20, surgicalness 1.0

[github-actions]

KernelGuard Blue Evaluation

• Evaluation job: 196
• Blue submission: 168
• Result: passed
• Target red submission caught: True
• Validation suite: TP 20/20, FP 20/20
• Surgicalness score: 1.0
• Workflow run: https://github.com/SinatrasC/kernelguard/actions/runs/25383080223

[SinatrasC]

Thanks for the KernelGuard Flywheel Campaign contribution. This PR is now superseded by the consolidated rule-family implementation in #273, which folds this detector coverage together with the related passing-eval variants.